DESCRIPTION: Recognition memory is a simple form of memory allowing well controlled experimental investigation. In the prototypical experiment, two separate presentations are made of a stimulus. In the first presentation the stimulus is novel and in the second it is familiar. In the last 25 years many laboratories have performed this type of experiment, seeking evidence of mechanisms underlying recognition by recording the spike trains from neurons in high level visual and association cortex of primates. One prominent candidate mechanism for recognition memory has been readily recorded under such circumstances, a decrement in response amplitude specific to previously presented images, called here a stimulus-specific adaptation (SSA). This SSA is robust; it has been recorded in many laboratories and is clearly evident in both single units and population averages from many areas (inferotemporal, perirhinal, entorhinal, V4, prefrontal cortex). Except under special behavioral circumstances, this is the only candidate mechanism which has been found for recognition memory in primates, and it is widely hypothesized to be the mediator of recognition memory. It was thus a surprise when the applicant was recently able to eliminate SSA from the neural responses of units recorded in inferotemporal cortex, without a concomitant elimination of behavioral recognition. The adaptation was eliminated by a using a unique, partial split-brain preparation in which two completely separate routes were used to bring information to the recorded units. Despite this result, SSA is still the leading, and currently really the only, candidate mechanism for recognition memory. The best way to reconcile this result with the hypothesis that SSA is the mechanism underlying recognition memory is to suggest that while the adaptation could be eliminated in inferotemporal cortex without eliminating behavioral recognition, this could not occur one or two stages further into higher level association cortex. To test this hypothesis, a series of experiments is proposed to determine whether SSA and behavioral recognition can be dissociated in entorhinal, perirhinal, and/or ventral prefrontal cortex.